Skip to content

README.md

Latest commit

 

History

History
100 lines (75 loc) · 3.3 KB

README.md

File metadata and controls

100 lines (75 loc) · 3.3 KB
Raw

Robot to Robot Handover

Motion Planning for Robotic Arm Handover Tasks Using Sampling-Based Methods (paper)

Motion-planning experiments for two robotic arms collaborating on a handover task. The project supports a 2D planar environment from envs/env.txt and a CoppeliaSim scene from envs/R2R.ttt.

Overview

Overview

Robots

2D Robot 3D Robot

Setup

conda env create -f environment.yml
conda activate r2r-handover

For --problem vrep, start CoppeliaSim first, open envs/R2R.ttt, and run the scene before launching the Python script.

Usage

python run.py [options]

Common options:

Option Description
--method {rrt,prm} Planner to run. Default: rrt.
--problem PROBLEM Environment file path or vrep. Default: envs/env.txt.
--connect Use RRT-Connect.
--bidirection Use Bidirectional RRT-Connect.
--star Use RRT*.
--num_samples N Number of planner samples. Default: 5000.
--step_length X Planner step length. Default: 0.15.
--connect_prob X Goal bias probability. Default: 0.05.
--local_planner {line,rrt} PRM local planner. Default: line.
--radius X PRM connection radius or RRT* rewiring radius. Default: 2.0.
--start "x y" Override start workspace coordinate. Must be used with --goal.
--goal "x y" Override goal workspace coordinate. Must be used with --start.
--no_draw Skip interactive display.
--no_save Do not save 2D frames or assets/robot_motion.gif.

Examples:

python run.py --help
python run.py --problem envs/env.txt --method rrt --connect
python run.py --problem envs/env.txt --method prm --local_planner line --no_draw --no_save
python run.py --problem vrep --method rrt --bidirection

Benchmarks

The benchmark scripts are import-safe and only run when executed directly.

python experiments/experiment_2d.py --variants rrt connect bidirectional star prm --output_dir results/2d
python experiments/experiment.py --variants rrt connect bidirectional star prm --output_dir results/3d --trace_star --trace_samples 1000

Use --sample_counts 100 for a quick smoke test. experiments/experiment.py requires CoppeliaSim and the remote API scene.

After the CSV files are generated, run plot.ipynb to recreate:

assets/result-2d.png
assets/result-3d.png
assets/result-time component.png
assets/result-test&dist.png

Files

Path Purpose
run.py CLI entry point for RRT/PRM runs.
planners/ RRT, RRT-Connect, Bidirectional RRT-Connect, RRT*, and PRM.
envs/ Environment files, 2D environment code, collision checks, plotting, and CoppeliaSim wrapper/scene.
experiments/ 2D and CoppeliaSim benchmark scripts.
assets/ Generated or reference visual assets.
results/ Generated benchmark CSV files.
environment.yml Conda environment definition.
requirements.txt Pip dependency list kept for fallback installs.